联邦二元门控与服务器端视觉-语言推理用于监控异常分类
Federated Binary Gating with Server-Side Vision-Language Inference for Surveillance Anomaly Classification
浏览论文内容
中文总结 AI 辅助
提出联邦二元CNN门控与服务器端VLM推理的混合架构，在UCF-Crime上实现异常分类，平衡传输与性能，提升灵敏度。
中文摘要 AI 辅助
隐私敏感的监控系统本可从大型视觉-语言模型（VLM）中受益，但此类模型通常需要集中访问原始视频。在联邦学习环境中，这一挑战因非独立同分布（non-IID）的客户端数据而加剧，这些数据可能使直接的多类异常分类变得不稳定，尤其是对于稀有类别。我们提出了一种混合两阶段架构，将联邦二元卷积神经网络（CNN）门控与服务器端零样本VLM推理相结合。轻量级LiteCNN3D门控执行本地异常筛选，仅将标记的视频转发给Qwen3-VL-8B，后者将其分配到四个异常元类。我们在将UCF-Crime分组为五个粗粒度元类的数据集上评估了该设计，并在真实的三节点异构部署中实现了联邦阶段。在所研究的设置中，直接联邦多类训练崩溃，而所提出的分解在分类质量和原始视频传输之间取得了更好的权衡。使用固定阈值路由时，联邦混合流水线几乎保持了与其集中式CNN+VLM对应物相同的宏平均F1分数（F1-macro），同时将传输视频的比例降低至51.4%，尽管其代理宏接收者操作特征曲线下面积（ROC AUC）低于集中式混合系统。一个补充的灵敏度导向路由操作点将宏ROC AUC从0.673提高到0.692，并将假阴性率从29.3%降低到22.9%，但将F1-macro从0.503降至0.485，同时将传输率从51.4%增加到57.9%。这些结果表明，联邦更适合粗粒度的本地筛选，而路由规则可以调整，以权衡服务器端VLM的使用，换取更高的异常灵敏度。
英文摘要
Privacy-sensitive surveillance systems could benefit from large vision-language models (VLMs), but such models typically require centralized access to raw video. In federated learning settings, this challenge is amplified by non-independent and identically distributed (non-IID) client data, which can make direct multiclass anomaly classification unstable, especially for rare categories. We propose a hybrid two-stage architecture that combines a federated binary convolutional neural network (CNN) gate with server-side zero-shot VLM inference. The lightweight LiteCNN3D gate performs local anomaly screening and forwards only flagged videos to Qwen3-VL-8B, which assigns them to four anomaly metaclasses. We evaluate this design on UCF-Crime grouped into five coarse metaclasses and implement the federated stage in a real three-node heterogeneous deployment. In the studied setting, direct federated multiclass training collapses, whereas the proposed decomposition yields a better trade-off between classification quality and raw-video transmission. With fixed-threshold routing, the federated hybrid pipeline preserves nearly the same macro-averaged F1 score (F1-macro) as its centralized CNN+VLM counterpart while reducing the fraction of transmitted videos to 51.4%, although with a lower proxy macro receiver operating characteristic area under the curve (ROC AUC) than the centralized hybrid system. A complementary sensitivity-oriented routing operating point increases macro ROC AUC from 0.673 to 0.692 and reduces the false negative rate from 29.3% to 22.9%, but decreases F1-macro from 0.503 to 0.485 while increasing transmission from 51.4% to 57.9%. These results suggest that federation is better suited to coarse local screening, while routing rules can be adjusted to trade server-side VLM usage for higher anomaly sensitivity.
发表机构
- ESIEA（ESIEA（法国高等科学与工程学院））
- CY Cergy Paris University（CY塞尔吉-巴黎大学）
- CNRS（法国国家科学研究中心）
机构由 AI 辅助整理，请以论文原文为准。
